Stereophile | January 2017 Issue
- Thread starter NorthStar
- Start date
This is what Ralph Karsten said about 2nd vs. 3rd - and I would agree its not so simple.
"Actually I am not a fan of lots of 2nd harmonic. And our amps really don't have any on account of their being fully differential- the 2nd is cancelled out at each stage in the amp. So this is a case where the topology can be used to control the distortion signature- we tend to get mostly the 3rd harmonic, and do as much as we can to keep it down, as we do with the output impedance. Is that a Kool-Aid? Dunno. It is certainly an attempt to keep known-distortion sources (pentodes, transistors, transformers) to a minimum, in order to reduce distortion. Some designers don't worry about distortion (SETs are typically 10% at full power, our amps are closer to 1%)."
We can all agree SET basically introduces loads of 2nd harmonic - whether this is audible or objectionable is up to the listener. People who like OTLs probably don't like SETs and vice versa.
"Actually I am not a fan of lots of 2nd harmonic. And our amps really don't have any on account of their being fully differential- the 2nd is cancelled out at each stage in the amp. So this is a case where the topology can be used to control the distortion signature- we tend to get mostly the 3rd harmonic, and do as much as we can to keep it down, as we do with the output impedance. Is that a Kool-Aid? Dunno. It is certainly an attempt to keep known-distortion sources (pentodes, transistors, transformers) to a minimum, in order to reduce distortion. Some designers don't worry about distortion (SETs are typically 10% at full power, our amps are closer to 1%)."
We can all agree SET basically introduces loads of 2nd harmonic - whether this is audible or objectionable is up to the listener. People who like OTLs probably don't like SETs and vice versa.
Driving a 83dB/w speaker having a low dip in impedance in the zone between 1 and 3 kHz and a 30º phase angle was not an wise decision ...
As most solid state amplifiers Gryphon's have plenty of feedback - they only claim zero global negative feedback.
We should not be afraid that it sounds different from SET's!
absolute power corrupts absolutely.
I love to be corrupted.
High power amplifiers usually have much higher capacitors, special care to isolate the input/driver circuits from the output stage, avoiding coupling through the supplies, thicker copper tracks and wires, more heavy and robust mechanics - all these aspects can result in a better sounding amplifier. Designers could easily have all these features in a model with lower voltage rails with lower power, but the marketing departments do not appreciate a model costing five times more and having half the power!
yes; more robust, heroic efforts to optimize the little things.
some very high power amplifiers have zero global feedback, along with minimal parts count in the signal path.....and so can still get that magical 1st watt.
getting musical realism is about choosing the right balance and the right compromises.....in amps and systems.
absolute power corrupts absolutely.
I love to be corrupted.
I agree....the more headroom the better. I love when the music soars like an Eagle.
Do you mean class B here or class AB? The latter was invented to deal specifically with this issue.
All this says is there are disagreements between designers and highly educated audiophiles. Time to go back to listening and hear a SET cannot drive anything less than ultra efficient speakers because a few posts so claim
I don't know.
OTOH it seems plausible that, even if a human can discern them equally that higher levels of any euphonic harmonic may not only discerned but preferred and used as a buying choice, whereas any harsh harmonic will cause rejection even at low levels.
Interesting post.
1. I am familiar with the distortion shortcomings of loudspeakers, that is why, IMHO, one notices much bigger sound differences between speakers than electronics in most hifi systems of my experience. It is very expensive and rare to produce low distortion in the bass, which is IMO very important since all the harmonics are in the audible range. Using modern design techniques to produce a linear magnetic circuit, sound engineering material choice and good structural design it is possible to produce a mid range driver with less than 0.1% distortion over its passband. There are more than 1 available and one that I know of is even very inexpensive. It is very expensive, again, to produce low distortion in treble units.
2. Do horns and panels produce 2nd harmonic too, or just conventionally coned traditional drivers?
3. I would expect the higher levels to be the more influential (though not totally dominant). In any case, if you are using the word "magnitude" in its normal scientific meanong of "x10" then you are not correct, very few speakers today have 10x more distortion than the amp driving them, 0.3% is not uncommon through the mid range nowadays. In fact the distortion of speakers is the same magnitude or less than SETs through the mid band nowadays.
4. I would expect so, but I would like to listen to that
OTOH 100Hz is lower than the lowest note produced by over half of orchestral instruments, so I am not so sure that distortion at that frequency is crucial to the reprouction of instrumental timbre, maybe, maybe not.5. We are not talking about vast enough changes and in recognising something is not "blabber" it is relative changes that count.
In my music classes in the 1960s we used to listen to exerpts and interpretations on a portable gramaphone which I am sure had several percent distortion and almost no bass but it was possible to discern the difference between an oboe and cor Anglais, despite them being very similar, presumably because the harmonics were changed identically so the differences remained.
6. Here you have me stumped. I have followed, or already seen, everything technical that other posters have mentioned until this. If the amplifier 2nd harmonic distortion was identical in magnitude to the speaker 2nd harmonic distortion over the entire audible band but of opposite phase they would indeed cancel, this is essentially how noise cancelling headphones and sports car ICE systems work. I am not aware of any known mechanism by which this cancellation could possible produce an artefact at a non-harmonic (or any actually) frequency such as the 3rd harmonic. How does this take place, please? So the logic of the idea that changing speaker phase is evidence that this is not an issue is false as far as I can see.
cheers
Thanks for taking the time
1.0.1% at what frequency and especially what level?85db? What about 95-100db.We are talking about realistic listening levels.
Distortions sub 500Hz and above 5.000Hz are above 1% even at ~90db.
2.Most drivers produce 2nd as dominant,panels especially their ribbons and the non-push pull panel motors.In horns,depends on the horn loading as well.
3.As i mentioned,speaker distortions are >1% below 500 and above 5000 at realistic listining levels.A good set will have 0.1% within .So yes,a magnitude higher in harmonic content will have a dominant effect(but of course not completely)
4.Sub 500Hz is,IMHO the base of music,orchestral or not,and this is where the majority of speakers start to fail.And as correctly stated,their hp will be all over the midband
5.The point is that 0.1% H2 will not change the timber of any instrument.I will agree that relevant changes is what matter.
6.Even harmonic cancellation is mathematically predictable,and measured.A chain of two components,either that is pre and amp,amp and speakers,even gain stages of an amp,will cancel H2 if in phase, and produce odd harmonics(usually higher than H3).
SO, here is the question.If a set sounds euphonic because of H2(hypothesis),and as the speaker will have H2 as dominant distortion profile(fact), the SE amplifier as dominant H2(fact) ,would the amplifier sound non euphonic if connected out of phase?(H2 cancelled)
It does not.
Cheers
Stavros
1.0.1% at what frequency and especially what level?85db? What about 95-100db.We are talking about realistic listening levels.
Distortions sub 500Hz and above 5.000Hz are above 1% even at ~90db.
2.Most drivers produce 2nd as dominant,panels especially their ribbons and the non-push pull panel motors.In horns,depends on the horn loading as well.
3.As i mentioned,speaker distortions are >1% below 500 and above 5000 at realistic listining levels.A good set will have 0.1% within .So yes,a magnitude higher in harmonic content will have a dominant effect(but of course not completely)
4.Sub 500Hz is,IMHO the base of music,orchestral or not,and this is where the majority of speakers start to fail.And as correctly stated,their hp will be all over the midband
5.The point is that 0.1% H2 will not change the timber of any instrument.I will agree that relevant changes is what matter.
6.Even harmonic cancellation is mathematically predictable,and measured.A chain of two components,either that is pre and amp,amp and speakers,even gain stages of an amp,will cancel H2 if in phase, and produce odd harmonics(usually higher than H3).
SO, here is the question.If a set sounds euphonic because of H2(hypothesis),and as the speaker will have H2 as dominant distortion profile(fact), the SE amplifier as dominant H2(fact) ,would the amplifier sound non euphonic if connected out of phase?(H2 cancelled)
It does not.
Cheers
Stavros
From Audio Electronics by John Linsley Hood (1999)
IMHO we must be very careful when debating distortions - what are the thresholds that separates "inaudible" from "euphonic" or "harsh"? Again IMHO we can not separate qualitative from quantitative opinions in this subject.
Attachments
Hi Stavros,
this could go on like ping-pong!
1. 0.1% is possible from ~250Hz to 1500Hz up to realistic listening levels. Not common but available.
2. My concern about my horns has always been more about sound reflections within the horn, very similar to the tuning of the induction and exhaust of racing engines
I have never seen a clean waterfall plot from horns, for example.
3. A good wideband speaker will not have distortion >1% as high as 500Hz simple 2 ways with small main units, maybe. A good 3-way with well engineered bass unit(s) go much lower than that. A quick look at the measurements of the well engineered but modestly priced (£1000) KEF R300 shows 0.7% @ 100Hz, 0.1% @ 1kHz and 0.1% at 10kHz at 90db, which is a realistic sound level IME. At almost double this price the Revel F36 has 0.4% @ 100Hz, 0.2% @ 1kHz and 0.1% @ 10kHz. Something mega like the Wilson Alexx measures 0.1% @ 100Hz. So I still don't agree with you about the speaker having a magnitude more distortion than the amp nowadays.
4. Because the speakers in 3 above achieve the performance they do I think you are exaggerating the speaker contribution at 500Hz. It is still contributing, yes, but not dominating the contribution of an SET type amplifier to the degree you suggest.
5. 0.1% will change the timbre IMO but it will change the timbre of all instruments similarly, so the instruments will still sound relatively similar but the individual timbre of each instrument will have had its timbre slightly changed.
6. In order to the 2nd harmonics to cancel they not only need to be out of phase but also of equal magnitude over the whole frequency range. This is highly unlikely/impossible from 2 different pieces of equipment so I do not believe connecting them out of phase will cause complete cancellation so the logic of your point here fails, the fact that they do not cancel proves nothing, and is even what should be expected.
I don't doubt that your amps sound lovely btw, I have heard many superb reports that they do!
cheers,
Frank
Hi Stavros,
this could go on like ping-pong!
1. 0.1% is possible from ~250Hz to 1500Hz up to realistic listening levels. Not common but available.
2. My concern about my horns has always been more about sound reflections within the horn, very similar to the tuning of the induction and exhaust of racing engines
I have never seen a clean waterfall plot from horns, for example.
3. A good wideband speaker will not have distortion >1% as high as 500Hz simple 2 ways with small main units, maybe. A good 3-way with well engineered bass unit(s) go much lower than that. A quick look at the measurements of the well engineered but modestly priced (£1000) KEF R300 shows 0.7% @ 100Hz, 0.1% @ 1kHz and 0.1% at 10kHz at 90db, which is a realistic sound level IME. At almost double this price the Revel F36 has 0.4% @ 100Hz, 0.2% @ 1kHz and 0.1% @ 10kHz. Something mega like the Wilson Alexx measures 0.1% @ 100Hz. So I still don't agree with you about the speaker having a magnitude more distortion than the amp nowadays.
4. Because the speakers in 3 above achieve the performance they do I think you are exaggerating the speaker contribution at 500Hz. It is still contributing, yes, but not dominating the contribution of an SET type amplifier to the degree you suggest.
5. 0.1% will change the timbre IMO but it will change the timbre of all instruments similarly, so the instruments will still sound relatively similar but the individual timbre of each instrument will have had its timbre slightly changed.
6. In order to the 2nd harmonics to cancel they not only need to be out of phase but also of equal magnitude over the whole frequency range. This is highly unlikely/impossible from 2 different pieces of equipment so I do not believe connecting them out of phase will cause complete cancellation so the logic of your point here fails, the fact that they do not cancel proves nothing, and is even what should be expected.
I don't doubt that your amps sound lovely btw, I have heard many superb reports that they do!
cheers,
Frank
Hello Frank
time is gold,and i respect my client's time, and yours, to pursue a ping pong game at the forum,so my last ball return would be below.
1.Have you had any real world distortion measurements of speakers?You will surprised to know that a 6db increase in output does not increase distortions by 6db. You will be surprised how fast distortions pick up after 95db.
Realistic listening levels are just that.Realistic.No room for opinions unfortunately,even acoustic un-amplified music peaks over 100db easy.That level should be at listening position.So add another 12db for 4m meter distance(6db for 2m) ,half of that for true line sources.
Have you seen real world distortion measurements at that levels?
10% for bass drivers,while tweeters ....well
Speakers do have one or two levels of magnitude higher THD when we are talking about realistic sound pressure.90db peaks is not realistic.Even vocals way surpass 90db.
2.Just because you haven't seen one it does not mean it does not exist.(what horns did you measure?)Commercial horns suffer from bad horn construction(resonances) ,poor flare choices(HOMS),bad termination(diffraction) among others.These will result in bad "waterfall" measurements.A proper horn can eliminate these problems..I will use an analogy you will appreciate more,if i spend my life driving cars with leaf suspension,it does not mean that racing suspensions do not exist.
(BTW i can show you horn-driver combo distortions , of -70db H2, at SP levels that most drivers will self desctruct.However, It does not mean that your average 100K floorstander will not have it's drivers jumping out of their magnetic gap at 105db
3.See 1
4.See 1
5.I am not arguing that,which is correct!I am arguing that 0.1% is audible,and various tests prove it is not.
6.If the amp and speaker have comparable H2 levels,and their distortion increase is monotonic,there will be a cancellation.No full ,but at least -20db.
One hour ago you did not know H2 cancellation existed and now you disregard it.I have measure it.Many have.It does happen.And it does not make an SE sound like a PP ,or the opposite...Food for thought
Again in a friendly tone.
Thank you for the compliment Frank,much appreciated.
Best
Stavros
Thanks for the reply, I am retired now so my time is very much less valuable than it was. When I was still working I never looked at sites like this.
I take on board all your points.
Personally I mostly listen at lower levels than I measure in concerts (I have measured both). Indeed a good solo soprano can be 90dB in a small concert hall, I was probably 15m away when I measured this and was very surprised.
I am about 6m from my Tune Audio Animas whilst listening. Sometimes I indulge myself with full concert loudness, I measure 105-110 dB peaks, so obviously much louder at 1m. Mostly I do not listen as loud as this.
I rarely listen to pop music.
I exaggerated about the waterfall plot of horns sorry, I have only ever seen two and the one which sticks in my mind is that of the Anima from the French review linked on the Tune Audio web site, I bought them anyway since they sound fabulous to me
It was not the ability to cancel using out of phase signals I wasn't aware of, the road car side were experimenting with this when I was Technical Director of Lotus 25 years ago, what I wasn't aware of is how such a cancellation could produce an odd harmonic as an artefact and this point is still unexplained but don't waste any time on it please.
thanks for your information, much appreciated.
Years since I last saw this but it is the evidence that listeners find 0.5% added 2nd harmonic distortion to be "more musical" which leads to the idea that people may well chose a piece of kit which adds this sort of distortion because they prefer it.
Again "does it sound good because of the distortion or despite it?" question, and again I firmly believe the former.
Just agreeing with those 1999 observations.Still do. Nobody has yet produced a cogent reason why people buy these items other than this.
For me it is no problem to buy what you like to listen to best, that makes sense.
If you want to justify it to yourself any way you like, fine.
So you have the Animas?Very nice. We got best of show with the excellent Animas in Munich's 2011 show paired with our full electronic set up.With our Exsequor amps.SE though